WO2011038989A1

WO2011038989A1 - Method and device for analyzing amplitude modulated broadband noise

Info

Publication number: WO2011038989A1
Application number: PCT/EP2010/062036
Authority: WO
Inventors: Heiko Schmidt; Jan-Philip Schwarz
Original assignee: Atlas Elektronik Gmbh
Priority date: 2009-09-29
Filing date: 2010-08-18
Publication date: 2011-04-07
Also published as: EP2483705A1; KR20120065408A; KR101330928B1; DE102009043417A1; EP2483705B1

Abstract

The invention relates to a method for analyzing amplitude modulated broadband noises emitted by bodies, in particular operating noises emitted by propeller driven vehicles, which are transferred as acoustic signals after a pre-processing 11 as input signals to an analysis device 1 wherein they are filtered by means of a band pass 3 according to the settings of the analysis device 1, demodulated by means of a demodulator 5 in respect of their envelopes, released from a DC component by means of a high pass 7 and subjected to a spectral analysis 9. The analysis device 1 generates so-called DEMON spectra, on the basis of which an evaluation of the settings of the analysis device 1 are carried out in an evaluation device 15. The filter settings of the band pass 3 are then changed by means of an adaptive band limit module 17 in dependence on a previously determined evaluation criterion. The invention furthermore relates to a device for carrying out said method.

Description

Method and apparatus for analyzing amplitude modulated broadband noise

The invention relates to a method for analyzing amplitude-modulated broadband noises radiated from bodies, in particular from operating noise radiated by propeller-driven vehicles, which are received as acoustic signals via a group of electroacoustic and / or optoacoustic transducers with a downstream directional generator, digitized and as input signals of a Analysis device to be handed over referred to in the preamble of claim 1. Art. Furthermore, the invention relates to an apparatus for carrying out the method according to claim 6.

It is known from airborne sound technology and also from waterborne sound technology that bodies can be detected and located on the basis of the operating noise radiated by them. For example, vehicles, particularly propeller-driven vehicles such as surface ships, manned or unmanned underwater vehicles, underwater hoppers, propeller aircraft or helicopters, generate operating noise during a ride which generates periodic volume fluctuations. These volume fluctuations occur only when the noise source not only radiates stochastic sound signals, but these sound signals are caused by drive units and rotating parts, such as propellers, which also include propellers. They are comparable to an amplitude modulation of a noise carrier and characteristic of any propeller driven vehicle.

Cavitation noise occurs in the water at a certain speed of the propeller blade. Cavitation refers to the formation and dissolution of cavities in liquids due to pressure fluctuations. The result of the propeller blades The resulting negative pressure is so great that gas bubbles are created whose implosion produce short acoustic pulses with great intensity. By virtue of their rotation in the water, cavitating propeller blades thus produce an amplitude-modulated broadband noise, in particular in the higher-frequency range. In order to obtain the information contained in the envelope of the modulated signal, conventionally the so-called DEMON analysis method is used (DEMON = Detection or Demodulation of Effects Modulation On Noise).

The DEMON analysis provides detailed information about a body based on its emitted acoustic signals. For example, if the radiated operating noise of a propeller-driven vehicle is present, the DEMON analysis provides information on the number of driving propellers, their number of revolutions and the number of blades, from which the type of vehicle can be deduced. It includes a bandpass to confine the received signals to the information carrying frequency domain, to demodulate and smooth a demodulator for the received signal with respect to its amplitude modulation, to remove a high pass from the DC component around the received signal, and spectral analysis to form so-called DEMON spectra the envelope of the received signals and for determining the fundamental frequencies and associated harmonics of spectral lines in the DEMON spectrum.

EP 2009 459 A1 shows a DEMON analysis method for the evaluation of underwater sound signals by passive sonar systems. In this case, the amplitude-modulated broadband noises resulting from, for example, ship propellers are filtered by means of a tree filter bank consisting of low and high passes, and thus divided into N individual subband signals. The subband signals determine the frequency ranges to be examined. The DEMON analysis always takes place here for each subband. The thus determined DEMON spectra are then evaluated separately, whereby a weighting function is determined which indicates how information-carrying the individual local areas. The overlay or combination as a "DEMON" DEMON spectrum is finally carried out by a suitably normalized intervals weighted overlay for each local area.

Conventionally, the desired demodulation band for DEMON analysis by means of the bandpass is set manually by an operator. This is done by changing the filter settings, in particular center frequency and bandwidth or upper and lower cutoff frequency and is possible only in a fixed grid. Further, there is no indication of optimal adjustment so far, forcing the operator to try all the settings. The disadvantage here is that the actual performance of the DEMON analysis essentially depends on whether a suitable demodulation band has been manually set via a corresponding operation of the bandpass.

The invention is therefore based on the problem of simplifying the selection of a demodulation band suitable for DEMON analysis.

The invention solves this problem by a method or a device having the features of claims 1 and 6, respectively. Via a receiving antenna of a sound system, water sound signals are received by means of a plurality of electroacoustic and / or optoacoustic transducers Water bodies have been radiated. These are preferably water sound signals with an amplitude modulation in the noise. Such signals arise, for example, by cavitation generating propeller blades of a propeller or other, periodic volume fluctuations generating drive types, such as a fin drive or a caterpillar drive. Such a sonar system also has a direction generator in which received signals generated in the transducers are processed to form group signals each having a directional characteristic. For this purpose, the received signals of the transducers are time-delayed in such a way that they are in relation to one another with respect to a main direction of the directional characteristic, and then added to form the group signal. The delay time around which the respective Receive signal is delayed, is determined by the position of the respective transducer with respect to the desired main direction of the directional characteristic. By differently sized time delays, a plurality of group signals and thus directional characteristics can be formed, which are pivoted to different sizes, in particular horizontal, angle against each other and against a reference plane, such as the normal plane of the receiving antenna.

For the DEMON analysis, the group signal of a directional characteristic, which is provided to an analysis device, is taken into account. For tasks such as detection, target separation and target tracking multi-channel processing is performed. Therefore, the DEMON analysis advantageously takes place in parallel for all directional characteristics or directional channels.

In order to simplify the selection of the demodulation band of the DEMON analysis, the settings of the analysis unit are evaluated, wherein for the evaluation, an evaluation criterion dependent on the settings of the analysis device is determined on the basis of the DEMON spectra. By means of an adaptive band limit module, the filter settings of the bandpass are controlled in dependence on the evaluation criterion in order to change the settings of the analysis device. The filter settings can be changed automatically or manually by an operator based on corresponding output values.

The adaptive bandlimit module alters the filter settings of the bandpass with the aim of maximizing the value of the rating criterion. This is done either by changing the center frequency and bandwidth of the filter or by changing the upper and lower cutoff frequencies.

The advantage of using an adaptive bandlimiting module is an improvement in detection performance and / or classification results. Further, automatic adjustment of the demodulation band improves and simplifies the separation and tracking of targets. Another advantage The method according to the invention provides the possibility of evaluating the settings of the analysis device. Thus, there is a measure of the quality or quality of the setting of the analyzer. This measure is advantageously used to find the optimum setting.

According to a preferred embodiment of the invention, the evaluation criterion is formed from a ratio of the strongest spectral line of the DEMON spectrum to a noise value of the DEMON spectrum. Advantageously, the DEMON spectra for this purpose are averaged over a predetermined time. The noise value of the DEMON spectrum is determined, for example, according to the median principle from all available spectral values or else as the mean value of all non-linear spectral components.

According to another embodiment of the invention, the evaluation criterion is formed from a ratio of a total level resulting from the sum of the levels of spectral lines of a previously determined line pattern of the DEMON spectrum to a noise value of the DEMON spectrum. This evaluation criterion is used as S / N value. The DEMON spectra are averaged over a predetermined time. The line patterns are formed by combining the strongest spectral lines with their harmonics. Subsequently, in each case a total level is calculated from the sum of the levels of the combined spectral lines. For the evaluation criterion, the maximum overall level is advantageously used.

According to a further embodiment of the invention, the change of the filter settings of the bandpass is carried out iteratively. For this purpose, the input signals of the analysis device are temporarily stored as a data block over a predetermined time in order to apply the analysis, the determination of the evaluation criterion and the changing of the filter settings of the bandpass filter several times to the same data block. This happens until a maximum of the evaluation criterion is reached. This method is particularly advantageous for single channel analysis for a directional channel. According to a further embodiment of the invention, the change of the filter settings of the bandpass filter is carried out continuously, whereby the determination of the evaluation criterion and the changing of the filter settings of the bandpass filter take place only once, so that the changes of the filter settings are applied to the subsequent analysis. The adaptive band limit module, starting from fixed start times, ensures that the filter settings are changed in small steps. If the change leads to an improvement of the evaluation criterion, the bandpass is adjusted accordingly. On the other hand, if the evaluation criterion worsens, the filter settings are varied in the opposite direction in the next analysis run. When choosing the step size for changing the filter settings, it should be noted that, for example due to target tracking, the best possible filter settings may change over time. If this is the case, then the step size should be at least so large that the adaptive band limit module can also follow the expected change. This method is particularly advantageous for multichannel analysis and improves the property of target separation and target tracking.

According to a further embodiment of the invention, the analysis of the amplitude-modulated broadband noises is carried out in one channel or in parallel for a plurality of directional channels generated by the direction generator. If the analysis is applied to individual directional channels in a fixed grid, the detection performance is advantageously improved. If the analysis is applied to controlled directional channels, the result of target separation and target tracking is advantageously improved. In this case, a directional channel is selected in the parallel, multi-channel processing, which provides the evaluation criterion available. This is either the direction channel with the maximum evaluation criterion or else the selection is controlled, for example via a destination tracking device. When tracking multiple targets, the analysis is performed on each target individually. Further advantageous embodiments will become apparent from the dependent claims and from the closer explained with reference to the accompanying drawings embodiments. In the drawing show:

Fig. 1 is a block diagram of the analysis of radiated operating noise, Fig. 2 shows a typical averaged DEMON spectrum.

The invention is described in more detail for the use of the method according to the invention in waterborne sound engineering in connection with a sonar system. However, the following explanations are also applicable to other methods such as airborne sounding methods.

Fig. 1 is a block diagram illustrating an analysis of amplitude modulated broadband noises which may be obtained, for example, from cavitating, i. Cavitation generating propeller blades propeller driven vehicles are produced, according to an embodiment of the invention.

However, the invention is not limited to propeller driven vehicles. All signals with an amplitude modulation in the noise can be examined by DEMON analysis. In this case, the signals may originate from bodies which lead to the formation of air bubbles, such as floats, paddle steamer or rowing boats, for example, because these bodies cause air bubbles on immersion of the arms or rudders or paddles due to their periodic movements in the water. Even divers can generate an amplitude-modulated noise with their breathing air during periodic inhalation and exhalation. Likewise, jetboats, ie small so-called personal watercraft, for example for up to four persons, generate amplitude-modulated broadband noise due to air bubble formation during the drive. Furthermore, other types of drive generate periodic volume fluctuations, such as a fin drive or a caterpillar drive, which also produces signals with an amplitude modulation in the noise. Shown is an analysis device 1, which has the conventionally known components such as bandpass 3, demodulator 5, high pass 7 and the spectral analysis 9.

Furthermore, the block diagram includes pre-processing means 1 1 in which the water sound signals are received by means of the transducers and electrical reception signals are generated. The preprocessing means comprise a direction generator which processes these received signals in a known manner into group signals of associated directional characteristics.

The group signal of a directional characteristic or a directional channel is passed to the analysis device 1 for performing a DEMON analysis. The group signal is filtered by means of the bandpass filter 3, rectified and smoothed by the demodulator 5 and freed from the DC component by means of the high-pass filter 7. Subsequently, the signal is subjected to spectral analysis 9 in order, after application of further imaging methods, to use BTR (BTR = Bearing Time Record), i. Waterfall representation of the bearing over time, or FRAZ (FRAZ = Frequency Azimuth Display), i. Representation of the frequency over azimuth to be represented. The spectral analysis 9 generates, for example by means of an FFT (Fast Fourier Transformation), a so-called DEMON spectrum, which is a frequency spectrum of the envelope of the received signal and emphasizes periodic modulation components. The result of the spectral analysis 9 is transmitted to both a target tracking device 13 and a rating device 15.

An evaluation of the setting of the analysis device 1 is carried out on the basis of the DEMON spectra by ascertaining in an evaluation device 15 an evaluation criterion resulting from the setting of the analysis device 1. This evaluation criterion is passed to an adaptive band boundary module 17.

In addition, the analysis includes a device, such as a switch 19, for selecting a manual change in the filter limits of the Bandpasses 3 by an operator 21 or an automatic change of the filter setting of the bandpass 3 due to the output values of the adaptive band boundary module 17th

FIG. 2 shows a typical averaged DEMON spectrum, as it appears after spectral analysis 9. On the horizontal axis 22, the frequency, preferably in Hertz, and plotted on the vertical axis 24, the corresponding level of the DEMON spectrum in decibels and the amplitude of the DEMON spectrum. There are clearly visible individual spectral lines 26 as lines in the background noise.

For the evaluation of the settings of the analysis device 1, the DEMON spectra are averaged over a predetermined period of time according to a so-called processing block length and passed to the evaluation device 15 for determining the evaluation criterion.

The evaluation is carried out, for example, by the observation of individual spectral lines 26. For this purpose, a ratio of the strongest spectral line to the noise level is determined. Such a determined ratio denotes the evaluation criterion and is used as S / N value. The noise level is determined, for example, according to the median principle from all present spectral values or else as the mean value of all non-linear spectral components.

Furthermore, it is conceivable to use whole line patterns for the evaluation of the setting of the analysis device. These are formed by combining the strongest spectral lines 26 with their harmonics. Subsequently, a total level is calculated in each case from the sum of the levels of the combined spectral lines. The ratio of the maximum total level to the noise level is used as the S / N value and corresponds to the evaluation criterion. The adaptive bandlimiting module 17 controls the filter settings of the bandpass filter 3 with the aim that the S / N value determined by the evaluation device 15 becomes maximum.

In an iterative change of the filter settings, the input signals of the analysis device 1 are temporarily stored as a data block for a predetermined time, so that the analysis, the determination of the evaluation criterion and the change of the filter settings can be performed multiple times on the same data block, namely, until a maximum of the S / N value of the evaluation device 15 or a maximum evaluation criterion is reached.

In the event of a continuous change in the filter settings, the determination of the evaluation criterion and the change of the filter settings are only applied once to the bandpass 3. This means that these changes only affect the subsequent input signals of the analyzer.

An exemplary procedure would be, starting from fixed start settings of the filter settings of the bandpass filter 3, to change both filter limits successively in small steps, for example by 100 Hz. Typical filter settings are, for example, 4 kHz to 6 kHz. The evaluation is carried out for each filter setting. If the change in the filter setting leads to an improvement in the evaluation criterion by increasing the S / N value, the bandpass filter 3 is adapted accordingly. However, if the rating criterion deteriorates by decreasing the S / N value, the filter setting is varied in the opposite direction in the next analysis run. In the choice of the step size, the frequency selectivity of the expected target signal and the selectivity of the bandpass 3 to be achieved are decisive. If it is to be expected that the best possible filter settings change over time, then the step size is to be chosen at least so large that the adaptive limiting module 17 can also follow the expected change. The changes in the filter settings of the bandpass filter 3 by means of the adaptive band limit module 17 are carried out as a function of the evaluation criterion determined by the evaluation device 15. The larger the determined S / N value or the evaluation criterion, the more suitable is the demodulation band set for the analysis of the noises. It is conceivable that the changes in the filter settings instead of automatically from the adaptive band limit module 17 make manually by an operator 21. This can be done as usual by trial and error or on the basis of correspondingly optimized output values of the adaptive band limit module 17.

The method according to the invention can be applied to individual directional channels in a fixed grid as well as to controlled directional channels. If there is a parallel, multichannel processing of the analysis, a direction channel is selected whose DEMON spectra are used to determine the evaluation criterion. This evaluation criterion is passed to the adaptive band limit module 17 for changing the filter settings of the bandpass 3. The selected direction channel is either the direction channel with the maximum rating criterion or else the selection is controlled, for example via the destination tracker 13. When tracking multiple destinations, it is necessary to perform the analysis for each destination individually.

With the method according to the invention, all noise sources can be analyzed in which an amplitude modulation occurs.

All mentioned in the foregoing description and in the claims features according to the invention both individually and in any combination with each other used. The invention is therefore not limited to the described or claimed feature combinations. Rather, all combinations of individual features are to be regarded as revealed.

Claims

claims

Method for analyzing broadband noise modulated by bodies, in particular operating noise emitted by propeller-driven vehicles, which are received as acoustic signals by means of a group of electroacoustic and / or optoacoustic transducers with a downstream directional generator, digitized and as input signals of an analysis device (1) in which they are filtered according to the settings of the analysis device (1) by means of a bandpass filter (3), demodulated by means of a demodulator (5) with respect to their envelope, freed from DC by means of a high-pass filter (7) and subjected to spectral analysis (9) which produces so-called DEMON spectra,

characterized in that

the settings of the analysis device (1) are evaluated, for the evaluation of an evaluation of the settings of the analyzer evaluation criterion is determined using the DEMON spectra and the settings of the analysis device are changed by controlling the filter settings of the bandpass (3) depending on the evaluation criterion by means of a adaptive band limit module (17).

Method according to claim 1,

characterized in that

the evaluation criterion is formed from a ratio of the strongest spectral line of the D E MO N spectrum to a noise value of the DEMON spectrum and this evaluation criterion is used as the S / N value.

Method according to claim 1 or 2,

characterized in that

the change of the filter settings of the bandpass filter (3) takes place iteratively, wherein the input signals of the analysis device (1) are temporarily stored as a data block over a predetermined time and thus the Analysis, the determination of the evaluation criterion and the changing of the filter settings is applied several times to the same data block and until such time as a maximum of the evaluation criterion is reached.

Method according to claim 1 or 2,

characterized in that

the change in the filter settings of the bandpass filter (3) is continuous, whereby the determination of the evaluation criterion and the changing of the filter settings occur only once, so that the changes of the filter settings are applied to the subsequent analysis.

Method according to one of the preceding claims,

characterized in that

the analysis is carried out in one channel or in parallel for a plurality of direction channels generated by the direction generator, wherein in the case of the parallel multi-channel processing, a direction channel is selected which provides the DEMON spectra for the determination of the evaluation criterion.

Device for analyzing broadband noises radiated by bodies, in particular of operating noise radiated by propeller-driven vehicles, receivable, digitizable and as input signals of an analysis device (1) as acoustic signals via a group of electroacoustic and / or optoacoustic transducers with a downstream directional device can be transferred in accordance with the settings of the analysis device (1) by means of a bandpass filter (3) demodulated by means of a demodulator (5) with respect to their envelopes, by a high pass (7) by a DC component and a spectral analysis (9 ), which produces so-called DEMON spectra, characterized by

a rating unit (15) for evaluating the settings of the analyzer (1) and determining one of the settings of the analyzer seeinrichtung dependent evaluation criterion on the basis of DEMO N spectra and

an adaptive band limit module (17) for changing the settings of the analysis device by controlling the filter settings of the bandpass filter (3) as a function of the evaluation criterion.

Device according to claim 6,

characterized in that

the evaluation criterion can be determined from a ratio of the strongest spectral line of the D E MO N spectrum to a noise value of the DEMON spectrum and this evaluation criterion can be used as S / N value.

Device according to one of claims 6 or 7,

marked by

an iterative calculation unit for iteratively changing the filter settings of the bandpass filter (3) by means of an adaptive bandlimit module (17) by buffering the input signals of the analysis device (1) as a data block and thus multiplying the analysis, the determination of the evaluation criterion and the changing of the filter settings are applicable to the same data block until a maximum of the evaluation criterion is reached.

Device according to one of claims 6 or 7,

marked by

a continuous calculation unit for continuously changing the filter settings of the bandpass filter (3) by means of an adaptive bandlimit module (17) by determining the evaluation criterion and changing the filter settings only once so that the changes in the filter settings are applicable to the subsequent analysis.

Device according to one of claims 6 to 9, marked by

a selection unit for selecting a direction channel generated by the direction generator, which in the case of single-channel or parallel multi-channel processing, provides the DEMON spectra for the determination of the evaluation criterion.